The missing middle in Thailand: the firm size distribution

Table 4-2 describes the firm size distribution using the LFS 1995 and 2005 and compares the estimates with others available from the literature. This comparison is intended as a plausibility check and in order to see how the distribution compares with those seen in developed countries. It should be noted that understanding the firm size distribution in Thailand is complicated by methodological inconsistencies between sources. Different government agencies use and report different firm size categories and employ different data collection methodologies (OECD, 2011). In particular, as noted above, the categories in the LFS do not overlap perfectly with the categories used in other sources, and we have no data on one-person or other microenterprises without paid workers. We should therefore bear in mind that the importance of microenterprises as sources of employment is understated.

Our results represent the situation for private sector employees. These were 29 and 37 percent of the active labour force in 1995 and 2005, respectively, but 79 and 81 percent of all wage workers.

Table 4-2: The firm size distribution in Thailand, percent of employment by firm size over time, estimates from LFS and from the literature

Category Thailand,

Sources: ^a Liedholm and Mead (1987), ^b Author’s computations based on LFS 1995, ^c Wiboonchutikula (2002), ^d Author’s computations based on LFS 2005 and, ^e OECD (2011).

The results clearly describe a dualistic distribution. For small and medium-sized firms, the percentage employed declines monotonically with firm size category, despite the fact that the ‘width’ of the categories increases. In 1995, 31 percent of private sector workers worked in firms with 9 or fewer employees and 28 percent in firms with between 10 and 49. Only 8 percent of workers were employed in medium-sized enterprises, with between 50 and 99 workers, but a substantial

114

proportion – 23 percent – worked in large firms, with over 100 employees. In the 2005 data we can distinguish firms with over 200 employees, which provides even more convincing evidence of the ‘missing middle’. Only 7 percent of workers are employed in the smaller group of large firms, with between 100 and 200 workers, while 23 percent work in the largest firms, of 200 workers or more.⁹⁶ When compared with developed countries (data on Germany and Japan from OECD, 2011, are provided as an example), this suggests a pronounced missing middle.

Contrary to the hypothesis in 4.2.3, the firm size distribution has not moved towards a log-normal distribution between 1995 and 2005. The proportion employed in the smallest firms has stayed roughly static, but there has been an increase of around 7 percent in the proportion working in the largest enterprises.

Our results differ from those of Wiboonchutikula (2002), who found, in 1996, a far lower share of employment in the smallest enterprises – just 6 percent worked in firms with 9 employees or fewer. Alongside this finding, the author found the share of large and medium-sized enterprises to be substantially higher than in most of the developing countries included in Tybout’s (2000) review. Firms with 50 or more employees accounted for 78 percent of jobs, and enterprises with between 10 and 50 workers 16 percent.

However, recent estimates by the OECD (2011) are in line with our results, as they also indicate a ‘missing middle’ pattern.⁹⁷ The categories do not overlap perfectly with those of the LFS, and in particular do not permit detailed analysis at the lower end of the firm size distribution, but the aggregated results are quite close to ours. The OECD (2011) report shows little significant difference between Thailand and OECD countries in terms of the proportion of employment, GDP and exports accounted for by SMEs (here defined as enterprises with fewer than 200 employees). These accounted for 77 percent of employment in 2005.

However, SMEs with more than 50 employees are underrepresented compared to OECD countries and to other Asian non-OECD countries, accounting for just 12 percent of employment.

The comparability of our findings to those of Liedlholm and Mead (1987) is limited, as they date to 1978. However, they also appear a plausible fit with our findings, and suggest a missing middle pattern. We therefore conclude that the have no employees. If these enterprises were to be included the pattern of the ‘missing middle’ would be expected to be even more pronounced.

96 In the analysis that follows we use the 1995 categories, for the sake of comparability.

97 Although using a different definition of ‘medium size’ than Tybout (2000).

115

Further insights come from examining the relationship between the firm size distribution and region (Table 4-3). The distribution of firm sizes in Bangkok is much more evenly spread than in other regions, closer to distributions observed in developed countries. In other words, it is not characterized by a ‘missing middle’.

In all of the other regions a large proportion of workers are employed in micro and small enterprises, with very few in either medium-sized firms. The Central region differs from the other regions as it does have a relatively large share, exceeding that of Bangkok, of workers in the largest firms. However, medium-sized firms are absent here too.

Over the period studied the ‘missing middle’ became more pronounced in all regions, with the share of workers in firms with between 20 and 99 workers falling across the board. The overall rise in the share of the largest enterprises can be traced mainly to the Central region, where it increased by almost 13 percentage points. The increase was somewhat smaller in Bangkok and the North and almost insignificant elsewhere. The share of the smallest microenterprises fell in all regions apart from the South, where it increased by over 10 percent.

Table 4-3: Firm size distribution by region, 1995 and 2005, percent

Bangkok Central Northern Northeastern Southern

Firm

size 1995 2005 1995 2005 1995 2005 1995 2005 1995 2005 1-4 14.15 13.40 23.65 16.97 31.17 26.57 29.49 25.62 34.20 44.91 4-9 11.09 9.45 17.81 14.05 25.94 24.45 25.49 27.14 21.22 15.88 10-19 12.73 12.27 11.54 10.66 18.39 19.52 15.37 18.79 14.15 11.68 20-49 18.90 17.60 8.35 7.89 10.20 10.02 9.98 9.87 11.15 9.15 50-99 13.35 12.19 7.32 6.15 4.79 4.31 6.02 4.05 6.85 5.89

>100 29.77 35.09 31.34 44.27 9.50 15.14 13.66 14.53 12.44 12.49 Source: Author’s computations based on LFS 1995 and 2005.

The firm distribution also varies by sector (results available from author upon request). MSEs are largely concentrated in the relatively poorly paid agricultural sector, and, to a lesser extent, services. Changes in the firm size distribution between 1995 and 2005 have served to accentuate this pattern. The proportion of large firms in industry (over 100 employees) grew significantly. It also increased slightly among service firms. In 2005, 45 percent of industrial workers and 23 percent of service sector employees worked in firms with 100 employees or more.

In other words, the character of these sectors changed significantly, from being made up mainly of smaller firms to being dominated by relatively large enterprises. This may be a reflection of changes in fast-growing industries such as automobiles and electronics. The change of firm size composition within the agricultural sector has been small.

116

It is also worth noting that firm size is strongly correlated with education, with larger firms particularly likely to hire more educated workers (Table 4-4). In 1995 the share of workers with post-secondary education was 14 percent in the largest firms, compared to just 2 percent in the smallest microenterprises. The share of workers with post-secondary education rose to about 25 percent in 2005 among the largest firms, but to just 3 percent in the smallest.

Table 4-4: Education level by firm size, 1995 and 2005, percent

1-4 5-9 10-19 20-49 50-100 >100 Total 1995

None 5.3 6.5 4.4 1.5 1.4 0.7 3.5

Less than elementary 48.0 48.8 45.0 36.4 25.1 23.0 38.7

Elementary 31.8 27.2 26.1 26.8 30.5 27.7 28.4

Lower secondary 9.3 10.1 11.6 9.3 15.2 18.1 12.3

Upper secondary 3.7 4.4 6.9 13.1 15.1 16.5 9.4

Post-Secondary 2.0 2.9 6.0 12.8 12.6 14.0 7.7

2005

None 6.0 5.0 5.1 3.5 1.4 0.9 3.6

Less than elementary 32.4 32.3 28.0 21.7 15.9 11.0 23.2

Elementary 32.1 31.1 29.5 23.4 21.9 18.0 25.8

Lower secondary 17.0 16.8 15.9 16.0 18.5 22.1 18.3

Upper secondary 9.2 9.5 10.7 15.4 17.9 22.8 14.8

Post-Secondary 3.4 5.3 10.7 20.0 24.4 25.1 14.3

Source: Author’s computations based on LFS 1995 and 2005

Overall, notwithstanding the problems of comparability between sources, we observe a clear ‘missing middle’ pattern in the firm size distribution in 1995 and 2005. We find this result broadly plausible in comparison with estimates from other sources. The ‘missing middle’ is visible in all regions except for Bangkok, where the distribution is close to the patterns observed in developed countries.

The LFS results indicate that, rather than declining as expected, the ‘missing middle’ has increased between 1995 and 2005 due to an increase in the proportion employed in the largest firms. This increase occurred mainly in the Central region, but also in Bangkok and the North of the country. Meanwhile, the share of medium-sized firms fell, slightly, across the board and the share of the smallest firms increased in the South.